AI Is Doing the Graduate Grunt Work. Here's What Employers Want From You Instead

For decades, the first year of a graduate job followed a familiar pattern. You learned the ropes by doing the routine work nobody senior wanted to do: reconciling spreadsheets, formatting reports, checking documents line by line. It was dull, but it was how you learned the job.

That model is changing fast, and some of the UK's biggest graduate employers are rebuilding their training around it.

What are the Big Four changing about graduate training?

In September 2026, City AM reported that KPMG, PwC, Deloitte and EY are all reshaping how they train graduates, because AI now handles much of the work first-year trainees used to do, including data entry, basic report preparation and manual processing (City AM).

The firms are responding in different ways. KPMG has moved professional exams earlier so graduates have the technical grounding before client work starts in their second year. PwC has introduced debating sessions. Deloitte and EY are putting more weight on communication, adaptability, problem-solving and critical reasoning. KPMG's chief people officer described the shift as graduates spending more time reviewing and exercising judgement rather than simply producing outputs.

The same report noted that the firms are reportedly taking on fewer graduates, with automation and a weaker economy both cited.

Is this only happening at big firms?

No, although big firms are moving fastest. A Work Foundation survey of 1,001 UK businesses, carried out in May 2026, found 36% had cut entry-level vacancies over the past year. Medium and large employers were more than twice as likely as small firms to say AI and automation were reducing entry-level roles, at around six in ten compared with a quarter (Work Foundation, via Global Business Outlook).

The direction is clear even if the pace varies. Fewer roles are being built around routine production, and more of the entry-level work that remains involves checking, interpreting and deciding.

What does judgement mean in an entry-level job?

Judgement sounds senior, but in a first job it is quite practical:

• Checking work, including AI's. Spotting the figure that doesn't add up, the reference that doesn't exist, the summary that misses the point.
• Prioritising. Knowing which of five requests matters most today, and saying so.
• Knowing what good looks like. Telling a strong draft from a weak one, and explaining why.
• Escalating well. Recognising when something needs a senior eye, and flagging it early with the facts.
• Explaining your reasoning. Showing how you reached a view, so other people can trust it.

None of these need years of experience. They need attention, curiosity and a willingness to question things.

How do you show judgement on your CV?

Most graduate CVs list tasks. To show judgement, write about the decisions inside those tasks:

• Instead of "Responsible for data entry for the sales team", write "Spotted and corrected duplicate customer records during weekly data entry, preventing double-counted sales in the monthly report."
• Instead of "Helped organise society events", write "Chose between three venues for a 200-person event by comparing cost, accessibility and transport links, then presented the recommendation to the committee."
• Instead of "Used AI tools for research", write "Used AI tools to draft literature summaries, then checked every source and removed references that didn't exist or didn't support the point."

Each version shows you noticed something, weighed it up and acted. That is exactly what employers are now hiring for.

Which interview questions test judgement?

Expect more questions that ask you to evaluate, not just describe. Common versions include:

• Tell me about a time you spotted a mistake before anyone else did.
• How would you check whether this report is accurate?
• Tell me about a decision you made without all the information.
• What would you do if two senior colleagues gave you conflicting priorities?
• Here is a piece of work. What would you change, and why?

Assessment centres test the same thing through case studies, in-tray exercises and group tasks where the right answer depends on what you choose to prioritise.

A strong answer walks through four steps: what you noticed, what you considered, what you decided and why, and what happened as a result. The "why" is the part most candidates skip, and it is the part assessors increasingly score. Our guide to how UK employers score interview answers goes deeper on this.

Should you use AI when you apply for jobs?

Many candidates do, so it is worth being straight about it. Using AI to draft, research or check your work is sensible, and it mirrors how you will work in the job. Letting it write answers you can't explain is a risk. If an interviewer asks you to expand on something in your application and you can't, the tool has cost you the job it helped you apply for.

A simple rule: use AI for the first draft and your own judgement for the final one. That is the same skill employers are looking for.

How can you practise showing judgement?

Judgement shows most clearly when you talk a decision through out loud, which is why practising spoken answers matters. MyPathfinder's Interview Simulator lets you rehearse with Elise, our AI interviewer, across 250+ industry-specific questions, with feedback after every answer. Practise the "why" until it comes naturally. If you want a structured way to practise before that, our guide to competency-based interview practice sets out a simple routine.

The bottom line

AI is removing the routine tasks that used to fill a graduate's first year, and the jobs that remain ask for judgement earlier. Show employers you can check, question, prioritise and explain, on paper and out loud, and you are applying for the job as it now exists rather than as it used to be.